Primary liver cancer is the sixth most frequent cancer globally and the second leading cause of cancer death. The most frequent liver cancer, accounting for approximately 85% of all primer malignant liver cancers and has a rising incidence, is hepatecellular carcinoma (HCC), which is formed by hepatocytes that become malignant. Another type of cancer formed by hepatocytes is hepatoblastoma, a rare malignant tumour that primarily develops in children, and accounts for approximately 1% of all cancers in children and 79% of all primary liver cancers under the age of 15. Secondary liver cancer, or metastasis in the liver, is a cancer that starts somewhere else in the body and then spreads to the liver. Examples of secondary liver cancer includes many common types of cancer, such as colon, rectum, lung, and breast cancer. Liver cancer can also form from other structures within the liver such as the bile duct, blood vessels and immune cells. Cancer of the bile duct (cholangiocarcinoma and cholangiocellular cystadenocarcinoma) account for approximately 6% of primary liver cancers.
While surgical resection and liver transplantation are potentially curative therapies for early stage HCC, more than 20% of the patients will eventually relapse or encounter further problems, and the majority of HCC diagnosis take place at a stage that is too advanced for these treatments. Regional therapies, such as radiofrequency ablation are a associated with response rates above 60%, but they are only suitable for a certain proportion of patients and are not always curative. Chemotherapy used so far has been minimally effective in HCC and to date response rates have not exceeded 25%. At present, sorafenib is the only effective drug on the market for the treatment of advanced or unresectable HCC, therefore, there is a great need for further treatments of HCC to reduce relapse rates and increase overall survival rates.
Many nucleoside analogues have been found to possess anticancer activity and they constitute a major class of chemotherapeutic agents that are widely used for the treatment of patients with cancer. This group of agents, known as antimetabolites, induces a variety of pyrimidine and purine nucleoside derivatives with cytotoxic activity.
Cellular nucleotide kinases phosphorylate nucleosides to their corresponding 5′-monophosphates which are further converted into their diphosphate and subsequently to the pharmacologically active triphosphate. It is known that some nucleosides are weakly active because they cannot be efficiently phosphorylated by kinases or are not substrates for kinases at all. In the phosphorylation sequence, the first phosphorylation of nucleoside analogues is rate limiting whereas the second and third phosphorylations are less sensitive to modifications to the nucleoside. Nucleoside monophosphates (nucleotides) per se are generally unstable in blood and show poor membrane permeation and hence are not suitable for use as drugs. Due to the high instability and poor cellular permeation of triphosphate of nucleosides and nucleoside analogues they cannot either be considered as possible drug candidates.
Troxacitabine, (beta-L-dioxolane cytidine) is a cytotoxic deoxycytidine analogue with an unnatural L-configuration which has demonstrated broad activity against both solid and hematopoietic malignancies in vitro and in vivo. Particularly, impressive activity has been observed against human cancer cell lines and xenografts ot hepatocellular, prostate, and renal origin (Cancer Res., 55, 3008-3011, 1995). Troxacitabine has shown to give rise to a mutation of the kinase deoxycytidine kinase (dCK) which is normally responsible for the first phosphorylation step of the nucleoside, leading to no or very low levels of troxacitabine monophosphate, thereby leading to resistance.
Troxacitabine entered phase III clinical trials in 2008 in the acute myelogenous leukemia indication, but did not proceed to registration. Discontinued phase II trials with troxacitabine include breast cancer, colorectal cancer, pancreatic cancer, melanoma, NSCLC, renal, prostate and ovarian tumours. Troxacitabine was generally administered as an intravenous infusion, thereby exposing many tissues to the drug, irrespective of the site of the cancer.
It has been shown that troxacitabine, despite its hydrophilic character, is transported into cells by passive diffusion, but is only very slowly accumulated in cancer cells in comparison with other, carrier transported nucleosides.
In WO2008/030373 derivatives of troxacitabine carrying a prodrug group on the cytosine base moiety are disclosed and the relationship between the lipophilicity of the prodrugs and their antitumor activity is evaluated. The patent states that base modification is desirable to avoid esterase difficulties with 5′—OH modification.
Phosphoramidate prodrugs at the 5′ hydroxyl function of D-nucleosides have been successfully employed in antiviral drugs, such as sofosbuvir used in the treatment of HCV infection.
Unmasking of the sofosbuvir prodrug to reveal the monophosphate intracetluiarly is a complex, multistep process involving several hydrolase enzymes in a particular sequence.
The use of phosphoramidate prodrugs on cancer nucleosides has been less successful. Nucana is developing Acelerin (Nuc-1031), a phosphoramidate prodrug of the D-nucleoside gemcitabine for the treatment of pancreatic cancer (for structure: see page 71 of WO2008012327). However, even though the phosphoramidate would be thought to enhance iipophilicity and cell permeability of the compound, the Acelarin prodrug must still be administered as an IV infusion, thus exposing many healthy tissues to the cytotoxic metabolite.
There is even less experience with monophosphate prodrugs of L-nucleosides such as troxacitabine. WO2008048128 discloses a small number of troxacitabine monosphosphate prodrugs including the compound at Example 14:

No cancer or other biological activity is disclosed for any of the compounds, either in the WO2008048128 specification or elsewhere in the academic literature. There are no reports of such a prodrug entering clinical trials. However, the inventors of WO2008048128 have published broadly similar prodrugs on the D-nucleoside gemcitabine (Baraniak et at Biorg Med Chem 2014 2133-2040) where the prodrug approach appears to work in certain tissues, and the D-nucleoside azidothymidine (Kulic et al Antivir Chem Chemother 2011 21(3) 143-150) where the prodrugs were 2-20 times less potent than the corresponding parent nucleoside. Kulic speculates that the azidothymidine prodrugs tended to be first dephosphorylated to the nucleoside and only then phosphorylated to the active triphosphate species. In that the prodrug approach works on gemcitabine (which resembles RNA by virtue of its substituted 2′ function), and does not work on azidothymidine (which is 2-deoxy thereby resembling DNA), it is hypothesised that the WO2008048128 prodrugs of troxacitabine (which is a DNA analog, albeit L-DNA), are likely to be inactive, like the azidothymidine prodrugs.
Balzarini et al Biochem Biophys Res Comm 225, 363-369 (1996) describe the HIV and HBV activity of CF 1109, a phosphoramidate prodrug of the L-nucleoside lamivudine/3TC, having the structure:

Balzarini states that this phosphoramidate prodrug was ˜250 fold less active against HIV than its parent nucleoside 3TC, but that the prodrug as “virtually equally effective against HBV in Hep G2.2.15 cells”. In other words, addition of this large, phosphoramidate methyl ester prodrug, group did not improve antiviral potency in a liver cell line. Balzarini did not assay whether the prodrug was being metabolized to 3TC prior to being phosphorylated to the active triphosphate.
The present invention provides phosphorus prodrugs of troxacitabine, particularly liver targeted prodrugs such as phosphoramidates, which are suitable for oral administration. These prodrugs have the advantage of improved cell permeability due to increased lipophilicity compared to troxacitabine per se, and to more efficient form the active triphosphate due to bypassing the rate limiting first phosphorylation step. Further, the compounds of the invention are primarily metabolised to the active triphosphate in the liver thereby providing a high concentration of active compound in the target organ and at the same time keeping side effects due to toxicity in other organs to a minimum.